World Bank Climate Data

Load and Clean Data

Load raw data

In the original dataset, NA value is labeled as “..”.

ClimateRaw1 = readxl::read_xls("Climate_Change.xls",sheet = 1,na = "..",col_types = "guess")
ClimateRaw2 = readxl::read_xls("Climate_Change.xls",sheet = 2,na = "..",col_types = "guess")
ClimateRaw3 = readxl::read_xls("Climate_Change.xls",sheet = 3,na = "..",col_types = "guess")

Visualize raw data

Cleaning data

When cleaning the datatable, there are several country code which is not actually country. Such as LIC represent low income countries and these rows represent an aggregated data. In order to separate these when plotting, we need to save these country codes for later use.
Pull out region and income codes

Regioncode = dplyr::filter(ClimateRaw2,Region == "Aggregates")[-grep("income",ClimateRaw2$`Country name`),] # Both the Income and Region aggregated rows will have a "Aggregates" value in the Region column
Regioncode = Regioncode %>% select(`Country code`) %>% pull()

Incomecode = dplyr::filter(ClimateRaw2,Region == "Aggregates")[grep("income",ClimateRaw2$`Country name`),] 
Incomecode = Incomecode %>% select(`Country code`)%>%pull()

There are 57 different variable observed for each country every year from 1990 to 2011(Ofcourse there are NA values,actually quite a lot. some is just missing, and some is measured every 10 years instead of every year.). After carefully examined 3rd sheet in the raw data, 9 variables are kept, they are:

Relatedvariables = dplyr::select(ClimateRaw3,`Series code`) %>% 
  slice(1:4,14:15,41:43)%>%
  pull()
print(Relatedvariables)

## [1] "SP.POP.TOTL"          "SP.POP.GROW"          "NY.GDP.MKTP.CD"      
## [4] "NY.GNP.PCAP.CD"       "SP.URB.TOTL"          "SP.URB.GROW"         
## [7] "EN.ATM.CO2E.PC"       "EN.ATM.CO2E.PP.GD.KD" "EN.ATM.CO2E.KT"

Total Population SP.POP.TOTL
Population anual growth SP.POP.GROW
GDP NY.GDP.MKTP.CD
GNI per Capita NY.GNI.PCAP.CD
Urban Population SP.URB.TOTL
Urban Population anual growth SP.URB.GROW
CO2 Emission per Capita EN.ATM.CO2E.PC
CO2 Emission per PPP EN.ATM.CO2E.PP.GD.KD
CO2 Total Emission EN.ATM.CO2E.KT

Using the codes to filter original Climate data

Climate <-  dplyr::filter(ClimateRaw1,`Series code` %in% Relatedvariables)
print(summary(as.factor(Climate$`Series code`)))

##       EN.ATM.CO2E.KT       EN.ATM.CO2E.PC EN.ATM.CO2E.PP.GD.KD 
##                  233                  233                  233 
##       NY.GDP.MKTP.CD       NY.GNP.PCAP.CD          SP.POP.GROW 
##                  233                  233                  232 
##          SP.POP.TOTL          SP.URB.GROW          SP.URB.TOTL 
##                  233                  232                  233

Because the second sheet of the original data contains categorical data indicating a country’s income status and geographic info. Assign these to the data table using dplyr::left_join, and also spread the table to become longer.

Climate <- left_join(Climate,select(ClimateRaw2,`Country code`,`Income group`,Region)) %>%
  melt(measure.vars = 7:28)%>%
  dplyr::rename(year = variable)

kable(Climate%>%dplyr::sample_n(10,replace = FALSE))%>%
  kable_styling() %>%
  scroll_box(width = "100%", height = "400px")

Country code	Country name	Series code	Series name	Decimals	Income group	Region	year	value
PYF	French Polynesia	SP.URB.GROW	Urban population growth (annual %)	1	High income: nonOECD	East Asia & Pacific	1996	1.594565e+00
WLD	World	SP.URB.TOTL	Urban population	0	Aggregates	Aggregates	2003	3.010250e+09
GRC	Greece	SP.POP.GROW	Population growth (annual %)	1	High income: OECD	Europe & Central Asia	2009	4.055627e-01
IDN	Indonesia	SP.POP.GROW	Population growth (annual %)	1	Lower middle income	East Asia & Pacific	2001	1.307325e+00
SAS	South Asia	EN.ATM.CO2E.KT	CO2 emissions, total (KtCO2)	1	Aggregates	Aggregates	1991	8.299118e+05
COG	Congo, Rep.	EN.ATM.CO2E.PC	CO2 emissions per capita (metric tons)	1	Lower middle income	Sub-Saharan Africa	1993	5.964629e-01
BEN	Benin	EN.ATM.CO2E.KT	CO2 emissions, total (KtCO2)	1	Low income	Sub-Saharan Africa	2005	2.566900e+03
TMP	Timor-Leste	SP.POP.TOTL	Population	0	Lower middle income	East Asia & Pacific	2008	1.079749e+06
BHR	Bahrain	SP.URB.TOTL	Urban population	0	High income: nonOECD	Middle East & North Africa	2011	NA
TUV	Tuvalu	EN.ATM.CO2E.PP.GD.KD	CO2 emissions per units of GDP (kg/$1,000 of 2005 PPP $)	1	Lower middle income	East Asia & Pacific	2007	NA

Get Series code table, Income code table as well as Region table

SeriesCode = ClimateRaw3%>%
  select(`Series code`,`Series name`,Definition)%>%
  dplyr::filter(`Series code` %in% Relatedvariables)

IncomeTable = ClimateRaw2%>%
  select(`Country code`,`Country name`)%>%
  filter(`Country code` %in% Incomecode)

IncomeTable = IncomeTable[c(1,6,5,3,2,4),]# re-order it

RegionTable = ClimateRaw2%>%
  select(`Country code`,`Country name`)%>%
  filter(`Country code` %in% Regioncode)

Check NA values in world data WLD, Region data and also Income group data

Climate%>%dplyr::filter(`Country code`=="WLD",is.na(value))%>%
  dplyr::group_by(year)%>%
  dplyr::select(`Series code`,year,value)%>%
  summary()

##  Series code             year       value    
##  Length:18          2011   :9   Min.   : NA  
##  Class :character   2008   :3   1st Qu.: NA  
##  Mode  :character   2009   :3   Median : NA  
##                     2010   :3   Mean   :NaN  
##                     1990   :0   3rd Qu.: NA  
##                     1991   :0   Max.   : NA  
##                     (Other):0   NA's   :18

Climate%>%dplyr::filter(`Country code`%in% Incomecode,is.na(value))%>%
  dplyr::group_by(year)%>%
  dplyr::select(`Series code`,year,value)%>%
  summary()

##  Series code             year        value    
##  Length:90          2011   :54   Min.   : NA  
##  Class :character   2009   :18   1st Qu.: NA  
##  Mode  :character   2010   :18   Median : NA  
##                     1990   : 0   Mean   :NaN  
##                     1991   : 0   3rd Qu.: NA  
##                     1992   : 0   Max.   : NA  
##                     (Other): 0   NA's   :90

Climate%>%dplyr::filter(`Country code`%in% Regioncode,is.na(value))%>%
  dplyr::group_by(year)%>%
  dplyr::select(`Series code`,year,value)%>%
  summary()

##  Series code             year        value    
##  Length:147         2011   :79   Min.   : NA  
##  Class :character   2010   :30   1st Qu.: NA  
##  Mode  :character   2009   :28   Median : NA  
##                     1990   : 4   Mean   :NaN  
##                     1991   : 3   3rd Qu.: NA  
##                     2008   : 3   Max.   : NA  
##                     (Other): 0   NA's   :147

As shown above, 2011 is missing too many datas, year 2011 will be excluded from the Dataset

Climate <- Climate%>%dplyr::filter(year != 2011)

check again

Climate%>%dplyr::filter(`Country code`=="WLD",is.na(value))%>%
  dplyr::group_by(year)%>%
  dplyr::select(`Series code`,year,value)%>%
  summary()

##  Series code             year       value    
##  Length:9           2008   :3   Min.   : NA  
##  Class :character   2009   :3   1st Qu.: NA  
##  Mode  :character   2010   :3   Median : NA  
##                     1990   :0   Mean   :NaN  
##                     1991   :0   3rd Qu.: NA  
##                     1992   :0   Max.   : NA  
##                     (Other):0   NA's   :9

Climate%>%dplyr::filter(`Country code`%in% Incomecode,is.na(value))%>%
  dplyr::group_by(year)%>%
  dplyr::select(`Series code`,year,value)%>%
  summary()

##  Series code             year        value    
##  Length:36          2009   :18   Min.   : NA  
##  Class :character   2010   :18   1st Qu.: NA  
##  Mode  :character   1990   : 0   Median : NA  
##                     1991   : 0   Mean   :NaN  
##                     1992   : 0   3rd Qu.: NA  
##                     1993   : 0   Max.   : NA  
##                     (Other): 0   NA's   :36

Climate%>%dplyr::filter(`Country code`%in% Regioncode,is.na(value))%>%
  dplyr::group_by(year)%>%
  dplyr::select(`Series code`,year,value)%>%
  summary()

##  Series code             year        value    
##  Length:68          2010   :30   Min.   : NA  
##  Class :character   2009   :28   1st Qu.: NA  
##  Mode  :character   1990   : 4   Median : NA  
##                     1991   : 3   Mean   :NaN  
##                     2008   : 3   3rd Qu.: NA  
##                     1992   : 0   Max.   : NA  
##                     (Other): 0   NA's   :68

Plot our data to have a taste of what it looks like

p <- ggplot(Climate%>%dplyr::filter(`Series code` == Relatedvariables[8],`Country code` %in% Regioncode))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+xlab("Year")+
  ggtitle(Relatedvariables[8])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -.3))

Found North America region is missing. In order to compare regions, we need to reconstruct data for North America.

dplyr::filter(ClimateRaw2,Region != "Aggregates")%>%select(Region)%>%pull()%>%unique()

## [1] "Latin America & Caribbean"  "Europe & Central Asia"     
## [3] "South Asia"                 "Sub-Saharan Africa"        
## [5] "Middle East & North Africa" "East Asia & Pacific"       
## [7] "North America"

From the dataset 3 countries are labeled in North America region: USA, CAN and BMU. Below are the reconstucting process.

# Check if NAS has been occupied
Climate%>%dplyr::filter(`Country code`=="NAS")

##  [1] Country code Country name Series code  Series name  SCALE       
##  [6] Decimals     Income group Region       year         value       
## <0 rows> (or 0-length row.names)

# Checked

Calculate Total CO2 emission, Total Population, GDP and Urban population

# Total CO2 emission
D1 <- Climate%>%
  dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[9])%>%
  dplyr::select(1:8)%>%
  dplyr::slice(1)

D1 <- D1%>%as.data.frame()

D1[1,c(1:2,7:8)] = c("NAS","North America","Aggregates","Aggregates")

D2 = Climate%>%
  dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[9])%>%summarise_at(.vars = 9:29,.funs = sum)
D = dplyr::bind_cols(D1,D2)

# Total Population
D1 <- Climate%>%
  dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[1])%>%
  dplyr::select(1:8)%>%
  dplyr::slice(1)

D1 <- D1%>%as.data.frame()

D1[1,c(1:2,7:8)] = c("NAS","North America","Aggregates","Aggregates")

D2 = Climate%>%
  dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[1])%>%
  summarise_at(.vars = 9:29,.funs = sum)
D <- dplyr::bind_rows(D, dplyr::bind_cols(D1,D2))

# GDP
D1 <- Climate%>%
  dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[3])%>%
  dplyr::select(1:8)%>%
  dplyr::slice(1)
D1 <- D1%>%as.data.frame()

D1[1,c(1:2,7:8)] = c("NAS","North America","Aggregates","Aggregates")

D2 = Climate%>%
  dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[3])%>%
  summarise_at(.vars = 9:29,.funs = sum)

D <- dplyr::bind_rows(D, dplyr::bind_cols(D1,D2))

# Urban Population
D1 <- Climate%>%dplyr::filter(Region == "North America")%>%pivot_wider(names_from = year,values_from = value)%>%dplyr::filter(`Series code` == Relatedvariables[5])%>%dplyr::select(1:8)%>%dplyr::slice(1)
D1 <- D1%>%as.data.frame()
D1[1,c(1:2,7:8)] = c("NAS","North America","Aggregates","Aggregates")
D2 = Climate%>%dplyr::filter(Region == "North America")%>%pivot_wider(names_from = year,values_from = value)%>%dplyr::filter(`Series code` == Relatedvariables[5])%>%summarise_at(.vars = 9:29,.funs = sum)
D <- dplyr::bind_rows(D, dplyr::bind_cols(D1,D2))

Calculate CO2 per capita, CO2 per unit PPP Bermuda’s PPP factor is missing from the world bank data, the cloest one I could find is a data given by federal reserve bank

pppBer <- 1.98021
pppCan <- 1.21
D_Temp <- D%>%dplyr::select(`Series name`,`1990`:`2010`)%>%
  column_to_rownames(var = "Series name")%>%
  transpose_df()
Temp_Colnames <- colnames(D_Temp)
colnames(D_Temp) <- c("year","CO2","Pop","GDP","Urban")
# Calculate ppp in 2005 for North america
ppp2005 <- Climate%>%dplyr::filter(Region == "North America")%>%
  pivot_wider(names_from = year,values_from = value)%>%
  dplyr::filter(`Series code` == Relatedvariables[3])%>%
  dplyr::select(`Country code`,`2005`)%>%
  column_to_rownames(var = "Country code")%>%
  transpose_df()%>%dplyr::transmute(`North America` = BMU*pppBer + CAN * pppCan + USA)%>%
  pull()

D_Temp = D_Temp%>%dplyr::mutate(CO2percap = CO2*1000/Pop,CO2perGDP = CO2*1000000000/ppp2005)

Adjust comparing variables exclude Pop growth and Urban growth and GNI per cap. (For simplicity)

SeriesCode <- SeriesCode%>%slice(-2,-4,-6)
# update related Variables
Related_variables <- SeriesCode%>%
  dplyr::select(`Series code`)%>%
  pull()

Temp_Colnames <- SeriesCode%>%
  dplyr::select(`Series name`)%>%
  pull()
#set adequate colname order and rename the D_Temp cols
colnames(D_Temp) <- c("year",Temp_Colnames[c(6,1:5)])
# Transpose back
D_Temp <- D_Temp%>%column_to_rownames(var = "year")%>%transpose_df()
colnames(D_Temp)[colnames(D_Temp)=="rowname"] <- "Series name"
print(colnames(D_Temp))

##  [1] "Series name" "1990"        "1991"        "1992"        "1993"       
##  [6] "1994"        "1995"        "1996"        "1997"        "1998"       
## [11] "1999"        "2000"        "2001"        "2002"        "2003"       
## [16] "2004"        "2005"        "2006"        "2007"        "2008"       
## [21] "2009"        "2010"

Reconstruct the remaining part of climate data

D_ = data.frame()

for (i in 1:length(Related_variables)){
  D1 <- Climate%>%
    dplyr::filter(Region == "North America")%>%
    pivot_wider(names_from = year,values_from = value)%>%
    dplyr::filter(`Series code` == Related_variables[i])%>%
    dplyr::select(1:8)%>%
    dplyr::slice(1)
  D1 <- D1%>%as.data.frame()
  D1[1,c(1:2,7:8)] = c("NAS","North America","Aggregates","Aggregates")
  D_ <- bind_rows(D_,D1)
}
# Combine D_ and D_Temp

D <- left_join(D_,D_Temp,by = "Series name")%>%pivot_longer(cols = 9:29,names_to = "year",values_to = "value")

Add these to the original Climate table and we keep the related variables

Climate <- bind_rows(Climate,D)
Climate <- Climate%>%
  dplyr::filter(`Series code` %in% Related_variables)
# update Regioncode
Regioncode <- append(Regioncode,"NAS")

Replot the previous plot

p <- ggplot(Climate%>%dplyr::filter(`Series code` == Related_variables[1],`Country code` %in% Regioncode))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+xlab("year")+
  ggtitle(Related_variables[1])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.3))

Series code is not user friendly, we would use Series name instead

SeriesName = SeriesCode%>%dplyr::select(`Series name`)%>%pull()

try plot

p <- ggplot(Climate%>%dplyr::filter(`Series name` == SeriesName[1],`Country code` %in% c(Incomecode,"WLD")))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+
  xlab("year")+
  ggtitle(SeriesName[1])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

p <- ggplot(Climate%>%dplyr::filter(`Series name` == SeriesName[2],`Country code` %in% c(Incomecode,"WLD")))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+
  xlab("year")+
  ggtitle(SeriesName[2])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

p <- ggplot(Climate%>%dplyr::filter(`Series name` == SeriesName[3],`Country code` %in% c(Incomecode,"WLD")))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+
  xlab("year")+
  ggtitle(SeriesName[3])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

p <- ggplot(Climate%>%dplyr::filter(`Series name` == SeriesName[4],`Country code` %in% c(Incomecode,"WLD")))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+
  xlab("year")+
  ggtitle(SeriesName[4])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

p <- ggplot(Climate%>%dplyr::filter(`Series name` == SeriesName[5],`Country code` %in% c(Incomecode,"WLD")))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+
  xlab("year")+
  ggtitle(SeriesName[5])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

p <- ggplot(Climate%>%dplyr::filter(`Series name` == SeriesName[6],`Country code` %in% c(Incomecode,"WLD")))+
  aes(x = as.numeric(as.character(year)),y = value,color = `Country name`)+
  geom_line()+
  xlab("year")+
  ggtitle(SeriesName[6])

plotly_build(p)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

We shall see how the data is distributed

TempD <- Climate%>%dplyr::filter(`Series name` == SeriesName[1],!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year %in% c(1992,2008))
p1 <- ggplot(TempD)+
  aes(x =log(value),color = `Region`,fill = `Region`)+facet_wrap(as.factor(TempD$year),nrow = 2)+
  geom_histogram(bins = 30,aes(y = ..density..),position = "dodge",alpha = .2)+
  geom_density(alpha = .3)+ylab("year")+
  xlab(paste0("log ",SeriesName[1]))

plotly_build(p1)#%>%layout(legend = list(orientation = 'h',x = 0, y = -0.2))

TempD <- Climate%>%dplyr::filter(`Series name` == SeriesName[2],!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year %in% c(1992,2008))
p1 <- ggplot(TempD)+
  aes(x =log(value),color = `Region`,fill = `Region`)+facet_wrap(as.factor(TempD$year),nrow = 2)+
  geom_histogram(bins = 30,aes(y = ..density..),position = "dodge",alpha = .2)+
  geom_density(alpha = .3)+ylab("year")+
  xlab(paste0("log ",SeriesName[2]))

plotly_build(p1)

TempD <- Climate%>%dplyr::filter(`Series name` == SeriesName[3],!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year %in% c(1992,2008))
p1 <- ggplot(TempD)+
  aes(x =log(value),color = `Region`,fill = `Region`)+facet_wrap(as.factor(TempD$year),nrow = 2)+
  geom_histogram(bins = 30,aes(y = ..density..),position = "dodge",alpha = .2)+
  geom_density(alpha = .3)+ylab("year")+
  xlab(paste0("log ",SeriesName[3]))

plotly_build(p1)

TempD <- Climate%>%dplyr::filter(`Series name` == SeriesName[4],!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year %in% c(1992,2008))
p1 <- ggplot(TempD)+
  aes(x =log(value),color = `Region`,fill = `Region`)+facet_wrap(as.factor(TempD$year),nrow = 2)+
  geom_histogram(bins = 30,aes(y = ..density..),position = "dodge",alpha = .2)+
  geom_density(alpha = .3)+ylab("year")+
  xlab(paste0("log ",SeriesName[4]))

plotly_build(p1)

TempD <- Climate%>%dplyr::filter(`Series name` == SeriesName[5],!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year %in% c(1992,2008))
p1 <- ggplot(TempD)+
  aes(x =log(value),color = `Region`,fill = `Region`)+facet_wrap(as.factor(TempD$year),nrow = 2)+
  geom_histogram(bins = 30,aes(y = ..density..),position = "dodge",alpha = .2)+
  geom_density(alpha = .3)+ylab("year")+
  xlab(paste0("log ",SeriesName[5]))

plotly_build(p1)

TempD <- Climate%>%dplyr::filter(`Series name` == SeriesName[6],!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year %in% c(1992,2008))
p1 <- ggplot(TempD)+
  aes(x =log(value),color = `Region`,fill = `Region`)+facet_wrap(as.factor(TempD$year),nrow = 2)+
  geom_histogram(bins = 30,aes(y = ..density..),position = "dodge",alpha = .2)+
  geom_density(alpha = .3)+ylab("year")+
  xlab(paste0("log ",SeriesName[6]))

plotly_build(p1)

try pivot series code,log all input

Climate1 <- Climate%>%
  dplyr::select(1:3,7:10)%>%
  pivot_wider(names_from = `Series code`,values_from = value)%>%
  dplyr::mutate_at(.vars = 6:11,.funs = log)

TempD <- Climate1%>%dplyr::filter(!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year ==2005 )

p3 <- ggplot(TempD)+aes(y = `EN.ATM.CO2E.KT`,x = `SP.POP.TOTL`)+geom_point(aes(size = `NY.GDP.MKTP.CD`,color = `Income group`))+scale_size_continuous(range = c(0,4))+xlab("Total Population, size = GDP")+ylab("Total CO2 Emission")

plotly_build(p3)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.3))

TempD <- Climate1%>%dplyr::filter(!`Country code` %in% Incomecode & !`Country code` %in% Regioncode,year ==2005 )

p3 <- ggplot(TempD)+aes(y = `EN.ATM.CO2E.PC`,x = `SP.POP.TOTL`)+geom_point(aes(size = `NY.GDP.MKTP.CD`,color = `Income group`))+scale_size_continuous(range = c(0,4))+xlab("Total Population, size = GDP")+ylab("Total CO2 Emission per Cap")

plotly_build(p3)%>%layout(legend = list(orientation = 'h',x = 0, y = -0.3))

Speculation: Higher GDP, Higher income is related to higher

Midterm Project

Weiling Li

10/20/2019

World Bank Climate Data

Load and Clean Data

Load raw data

Visualize raw data

Cleaning data